Anthropic: Claude Sonnet 5

Everyday Claude agent model for coding, planning, browsing, and general work

claude-sonnet-5

AvailableAnthropicChatCached input 90% less

Input / Output-70%: $2.00 / $10.00$0.60 / $3.00
Context: 1M
Released: Jun 30, 2026
Max output: 128K
Modalities: VisionChat
Capabilities: Tool usePrompt CacheReasoning

About Claude Sonnet 5

Claude Sonnet 5 is Anthropic's mid-tier model released on June 30, 2026, as the everyday agent model for coding, planning, and browsing. It narrows the gap with Opus 4.8 at a lower cost, finishes complex tasks where earlier models stopped short, and checks its own output without being asked. Sonnet 5.5 now replaces it as the current Sonnet.

Where it works well

Anthropic says it finishes complex tasks where predecessors would stop and checks its own work without being prompted.
One early tester described it writing a reproducing test, fixing the bug, then stashing the fix to confirm the test failed without it.
On BrowseComp it matches Opus 4.8 at medium effort, which makes it a cost-efficient agentic search model.
Adaptive thinking is on by default, so harder prompts get more reasoning without extra settings.

When to choose another model

Anthropic reports substantially weaker dangerous cyber skills than Opus 4.8, and it shows slightly higher misaligned behavior on behavioral audits.
Setting temperature, top_p, or top_k to non-default values returns an error.
Sonnet 5.5 is faster and stronger on Terminal-Bench 4.0, so new builds should start there.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Text to textMessages API

POST/v1/messages

API format:

curl https://api.tokenlab.sh/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: sk-xxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-5",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Default spec

per 1M tokens

Official price: Input $2.00 / Output $10.00 / Cache read $0.20 / Cache write $2.50
Verified price: Input $0.60 / Output $3.00 / Cache read $0.06 / Cache write $0.75
Discount: -70%

Prompt cache pricing

Cache read

Official price: $0.20
Verified price: $0.06
Discount: -70%

Cache write

Official price: $2.50
Verified price: $0.75
Discount: -70%

Tool fees

Charged per call, only when the model uses the tool.

Web search

Official price: $0.01/search
Verified price: $0.003/search
Discount: -70%

	Official priceper 1M tokens	Verified priceper 1M tokens	Discount
Default spec	Input $2.00 / Output $10.00 / Cache read $0.20 / Cache write $2.50	Input $0.60 / Output $3.00 / Cache read $0.06 / Cache write $0.75	-70%
Prompt cache pricing
Cache read	$0.20	$0.06	-70%
Cache write	$2.50	$0.75	-70%
Tool feesCharged per call, only when the model uses the tool.
Web search	$0.01/search	$0.003/search	-70%

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Oct 10, 2026, 12:25 PM UTC

30-day success rate: 99.7%
7-day median latency: 1.9 sn=275
7-day P95 latency: 6.8 sn=275
30-day requests: 100+
Last active: 3 days ago

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Claude Sonnet 5 in Console with a prompt ready to edit or send.

Help me try claude-sonnet-5 with a short message at /v1/messages. Show the reply, latency, and cost.

Use cases

Best for

Reasoning
Vision

Everyday coding agent
Run it in an editor or CLI agent to take issues from report to tested fix, with the model checking its own change.
Web research agent
Let it browse, compare sources, and compile a sourced answer to an open-ended question.
Planning and operations
Turn a goal into a plan, call internal tools to gather facts, and report back with open questions.

Prompt examples

Investigate why this endpoint returns 500 for empty carts, write a test that reproduces it, then fix it.

Find three independent sources on this vendor's outage history and summarize where they disagree.

Draft a rollout plan for the new pricing page, naming the owners, dependencies, and what could block launch.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What is Claude Sonnet 5 best for?

Everyday agent work: coding, planning, browsing, and general tasks at a lower cost than Opus. Anthropic says it narrows the gap with Opus 4.8 and is more likely to finish a hard task and check its own output.

How does Sonnet 5 differ from Sonnet 4.6?

Sonnet 5 is a newer generation with stronger tool use, coding, and knowledge work, adaptive thinking on by default, and a tokenizer that maps the same text to roughly 1.0 to 1.35 times more tokens. It also rejects non-default sampling parameters.

Should I move from Sonnet 5 to Sonnet 5.5?

Yes for new work. Anthropic reports Sonnet 5.5 generates output faster and uses fewer tokens per task, with much higher agentic coding scores. Expect breaking changes: forced tool use errors and thinking blocks tied to the model.

Does Claude Sonnet 5 read images?

Yes. It takes images along with text and returns text, and Anthropic reports improved computer-use results on OSWorld-Verified. It cannot produce images, so use a separate image model for that.

How much does Claude Sonnet 5 cost?

On TokenLab, Claude Sonnet 5 costs Input $0.60 / Output $3.00 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Claude Sonnet 5?

Claude Sonnet 5 accepts up to 1,000,000 tokens of context and returns up to 128,000 tokens in one response.

Which endpoint should Claude Sonnet 5 use?

Use https://api.tokenlab.sh/v1/messages for Claude Sonnet 5. The request example below shows the matching code shape.

Which operations does Claude Sonnet 5 support?

Claude Sonnet 5 supports Text to text. Select an operation above to see its endpoint and request example.